Business As Usual

My apologies to everyone for the site “outage” today. Some major kaboom happened with my provider, CrystalTech, and I had to convince them I wasn’t insane. The site came back up in the evening the same weird way it went down earlier in the day. Looks like I’m going to have to prove my sanity the second time.

Speaking of CrystalTech… I know—reviewing hosting companies is no easy thing. It’s enough to have only one really nasty experience with your host to hate them for the rest of your life. Today was surely the day they came close to pushing my patience over the edge.

CrystalTech has outstanding tech support—I’ve almost never had to wait to speak with someone, and you don’t get bounced to Bombay (as it happens with Dell). They reply to emails quite fast, follow up on open tickets, and there seem to be .NET devs around (at least I always spoke to one when I needed to) in case you’re seeing weird deployment issues.

The downside of this company is weak monitoring tools. I’ve had some “rogue process” pegging my server so bad that I couldn’t access my stats for days! They were surprised when I brought it to their attention… a hundredth time. I had to get mad to convince them.

Also, their SQL Server is very flaky. It’s quite often that I come in the morning and get a dozen or two of error alerts (custom error pages, eh?) that the SQL connection pool got exhausted, or SQL Server was shutting down and wasn’t accepting connections, etc. CrystalTech guys always swear none of this happens. I guess a whole exception stack trace is something by sick mind dreams up. It must be like a Magic Eye drawing—if you stare at a stack trace long enough, you’ll see a butterfly or a deer. :)

Modus operandi of CrystalTech is, “It’s not us. It’s all you.” Try to reproduce your issue, collect stack traces (if any), keep track of what you change and when, etc, and present all this to them. Otherwise you’ll be chasing your tail.